|
|
Created:
7 years, 6 months ago by petarj Modified:
7 years, 5 months ago CC:
native-client-reviews_googlegroups.com, ilevy-cc_chromium.org Base URL:
http://git.chromium.org/native_client/src/native_client.git@master Visibility:
Public. |
Description[MIPS] Switch the script to use packages from Debian Wheezy
Recent dependencies in Chromium (such as the need to use the latest libnss3
package) prevent us to use Debian Squeeze repository anymore.
This change switches it to Debian Wheezy repository and adds minor changes
in the script that ensure the toolchain build does not re-download tarballs
and SVN checkouts if they are already present.
The change also fixes issue #3401 when strcasecmp is statically linked into
the QEMU binary.
BUG= https://code.google.com/p/nativeclient/issues/detail?id=3401
BUG= https://code.google.com/p/nativeclient/issues/detail?id=2275
TEST= trusted-toolchain-creator.mipsel.debian.sh nacl_sdk (or chrome_sdk)
Committed on behalf of Petar Jovanovic (petarj@mips.com) because the
commit queue is currently not working for this change. (The presubmit
checks are timing out because they are invoking SVN, which is
prompting for credentials.)
Committed: https://src.chromium.org/viewvc/native_client?view=rev&revision=11674
Patch Set 1 #
Total comments: 10
Patch Set 2 : Code review comments. #Patch Set 3 : Add a comment. #
Messages
Total messages: 39 (0 generated)
PTAL.
On 2013/06/12 23:12:09, petarj wrote: > PTAL. Ping.
LGTM https://codereview.chromium.org/16880002/diff/1/tools/trusted_cross_toolchain... File tools/trusted_cross_toolchains/trusted-toolchain-creator.mipsel.wheezy.sh (left): https://codereview.chromium.org/16880002/diff/1/tools/trusted_cross_toolchain... tools/trusted_cross_toolchains/trusted-toolchain-creator.mipsel.wheezy.sh:619: --static Please add "BUG=https://code.google.com/p/nativeclient/issues/detail?id=3401" to the commit message and comment that this change fixes a bug. https://codereview.chromium.org/16880002/diff/1/tools/trusted_cross_toolchain... File tools/trusted_cross_toolchains/trusted-toolchain-creator.mipsel.wheezy.sh (right): https://codereview.chromium.org/16880002/diff/1/tools/trusted_cross_toolchain... tools/trusted_cross_toolchains/trusted-toolchain-creator.mipsel.wheezy.sh:1: #!/bin/bash Can you rename this file to remove "wheezy" from the name? It doesn't make much sense to have to rename it each time it switches to targeting a different Debian release. https://codereview.chromium.org/16880002/diff/1/tools/trusted_cross_toolchain... tools/trusted_cross_toolchains/trusted-toolchain-creator.mipsel.wheezy.sh:197: Did you intend to add this empty line? It seems superfluous. https://codereview.chromium.org/16880002/diff/1/tools/trusted_cross_toolchain... tools/trusted_cross_toolchains/trusted-toolchain-creator.mipsel.wheezy.sh:578: Banner "Fixing includes" Is this doing something that the Debian packages' postinst steps normally do? If so, please add a comment explaining that. https://codereview.chromium.org/16880002/diff/1/tools/trusted_cross_toolchain... tools/trusted_cross_toolchains/trusted-toolchain-creator.mipsel.wheezy.sh:624: readonly liblist="libbz2.so \ Make this '\' consistent with the following lines
PTAL. https://codereview.chromium.org/16880002/diff/1/tools/trusted_cross_toolchain... File tools/trusted_cross_toolchains/trusted-toolchain-creator.mipsel.wheezy.sh (left): https://codereview.chromium.org/16880002/diff/1/tools/trusted_cross_toolchain... tools/trusted_cross_toolchains/trusted-toolchain-creator.mipsel.wheezy.sh:619: --static On 2013/06/18 15:26:21, Mark Seaborn wrote: > Please add "BUG=https://code.google.com/p/nativeclient/issues/detail?id=3401" to > the commit message and comment that this change fixes a bug. Done. https://codereview.chromium.org/16880002/diff/1/tools/trusted_cross_toolchain... File tools/trusted_cross_toolchains/trusted-toolchain-creator.mipsel.wheezy.sh (right): https://codereview.chromium.org/16880002/diff/1/tools/trusted_cross_toolchain... tools/trusted_cross_toolchains/trusted-toolchain-creator.mipsel.wheezy.sh:197: On 2013/06/18 15:26:21, Mark Seaborn wrote: > Did you intend to add this empty line? It seems superfluous. Removed. https://codereview.chromium.org/16880002/diff/1/tools/trusted_cross_toolchain... tools/trusted_cross_toolchains/trusted-toolchain-creator.mipsel.wheezy.sh:578: Banner "Fixing includes" On 2013/06/18 15:26:21, Mark Seaborn wrote: > Is this doing something that the Debian packages' postinst steps normally do? > If so, please add a comment explaining that. No, it is not a postinst step. The Wheezy version has different locations for a few include files, and pkg-config does not seem to work correctly in some cases, so this is a workaround for it. https://codereview.chromium.org/16880002/diff/1/tools/trusted_cross_toolchain... tools/trusted_cross_toolchains/trusted-toolchain-creator.mipsel.wheezy.sh:624: readonly liblist="libbz2.so \ On 2013/06/18 15:26:21, Mark Seaborn wrote: > Make this '\' consistent with the following lines Done.
LGTM still https://codereview.chromium.org/16880002/diff/1/tools/trusted_cross_toolchain... File tools/trusted_cross_toolchains/trusted-toolchain-creator.mipsel.wheezy.sh (right): https://codereview.chromium.org/16880002/diff/1/tools/trusted_cross_toolchain... tools/trusted_cross_toolchains/trusted-toolchain-creator.mipsel.wheezy.sh:578: Banner "Fixing includes" On 2013/06/19 23:12:33, petarj wrote: > On 2013/06/18 15:26:21, Mark Seaborn wrote: > > Is this doing something that the Debian packages' postinst steps normally do? > > If so, please add a comment explaining that. > > No, it is not a postinst step. The Wheezy version has different locations > for a few include files, and pkg-config does not seem to work correctly in > some cases, so this is a workaround for it. OK, maybe you could add a comment that this is done because pkg-config is not working correctly?
CQ is trying da patch. Follow status at https://nativeclient-status.appspot.com/cq/petarj@mips.com/16880002/12001
Presubmit check for 16880002-12001 failed and returned exit status -2001. The presubmit check was hung. It took 360.0 seconds to execute and the time limit is 360.0 seconds.
On 2013/06/24 00:53:11, I haz the power (commit-bot) wrote: > Presubmit check for 16880002-12001 failed and returned exit status -2001. > The presubmit check was hung. It took 360.0 seconds to execute and the time > limit is 360.0 seconds. I will submit this one with NOTRY=true when try-bots finish the execution.
CQ is trying da patch. Follow status at https://nativeclient-status.appspot.com/cq/petarj@mips.com/16880002/12001
Presubmit check for 16880002-12001 failed and returned exit status -2001. The presubmit check was hung. It took 360.0 seconds to execute and the time limit is 360.0 seconds.
CQ is trying da patch. Follow status at https://nativeclient-status.appspot.com/cq/petarj@mips.com/16880002/12001
Presubmit check for 16880002-12001 failed and returned exit status -2001. The presubmit check was hung. It took 360.0 seconds to execute and the time limit is 360.0 seconds.
CQ is trying da patch. Follow status at https://nativeclient-status.appspot.com/cq/petarj@mips.com/16880002/12001
Checked... -BradN Sent from my phone. On Jun 24, 2013 6:46 AM, <commit-bot@chromium.org> wrote: > CQ is trying da patch. Follow status at > https://nativeclient-status.**appspot.com/cq/petarj@mips.** > com/16880002/12001<https://nativeclient-status.appspot.com/cq/petarj@mips.com/16880002/12001> > > > https://chromiumcodereview.**appspot.com/16880002/<https://chromiumcodereview... > -- You received this message because you are subscribed to the Google Groups "Native-Client-Reviews" group. To unsubscribe from this group and stop receiving emails from it, send an email to native-client-reviews+unsubscribe@googlegroups.com. To post to this group, send email to native-client-reviews@googlegroups.com. Visit this group at http://groups.google.com/group/native-client-reviews. For more options, visit https://groups.google.com/groups/opt_out.
Presubmit check for 16880002-12001 failed and returned exit status -2001. The presubmit check was hung. It took 360.0 seconds to execute and the time limit is 360.0 seconds.
CQ is trying da patch. Follow status at https://nativeclient-status.appspot.com/cq/petarj@mips.com/16880002/12001
Presubmit check for 16880002-12001 failed and returned exit status -2001. The presubmit check was hung. It took 360.0 seconds to execute and the time limit is 360.0 seconds.
CQ is trying da patch. Follow status at https://nativeclient-status.appspot.com/cq/petarj@mips.com/16880002/12001
Presubmit check for 16880002-12001 failed and returned exit status -2001. The presubmit check was hung. It took 360.0 seconds to execute and the time limit is 360.0 seconds.
CQ is trying da patch. Follow status at https://nativeclient-status.appspot.com/cq/petarj@mips.com/16880002/12001
Presubmit check for 16880002-12001 failed and returned exit status -2001. The presubmit check was hung. It took 360.0 seconds to execute and the time limit is 360.0 seconds.
As I commented on the other thread, the presubmit check is running a remote svn info command, which will not work on the CQ as it currently exists. The command hangs on a prompt for credentials. On Tue, Jun 25, 2013 at 10:51 AM, <commit-bot@chromium.org> wrote: > Presubmit check for 16880002-12001 failed and returned exit status -2001. > The presubmit check was hung. It took 360.0 seconds to execute and the time > limit is 360.0 seconds. > > > > https://chromiumcodereview.**appspot.com/16880002/<https://chromiumcodereview... > -- You received this message because you are subscribed to the Google Groups "Native-Client-Reviews" group. To unsubscribe from this group and stop receiving emails from it, send an email to native-client-reviews+unsubscribe@googlegroups.com. To post to this group, send email to native-client-reviews@googlegroups.com. Visit this group at http://groups.google.com/group/native-client-reviews. For more options, visit https://groups.google.com/groups/opt_out.
On 25 June 2013 10:58, Robert Iannucci <iannucci@chromium.org> wrote: > As I commented on the other thread, the presubmit check is running a > remote svn info command, which will not work on the CQ as it currently > exists. The command hangs on a prompt for credentials. > Which other thread was that? Would you happen to know what part of the code is calling "svn info", or do you only know this based on the prompt you're seeing? I had a quick look at NaCl's PRESUBMIT.py and tools/code_hygiene.py and didn't see any obvious svn invocations. BTW, does the CQ redirect stdin to /dev/null? Would that help make this fail fast rather than hang? Cheers, Mark > On Tue, Jun 25, 2013 at 10:51 AM, <commit-bot@chromium.org> wrote: > >> Presubmit check for 16880002-12001 failed and returned exit status -2001. >> The presubmit check was hung. It took 360.0 seconds to execute and the >> time >> limit is 360.0 seconds. >> >> https://chromiumcodereview.appspot.com/16880002/ >> > > -- You received this message because you are subscribed to the Google Groups "Native-Client-Reviews" group. To unsubscribe from this group and stop receiving emails from it, send an email to native-client-reviews+unsubscribe@googlegroups.com. To post to this group, send email to native-client-reviews@googlegroups.com. Visit this group at http://groups.google.com/group/native-client-reviews. For more options, visit https://groups.google.com/groups/opt_out.
I'm not sure, but I see presubmit_support.py with a svn info child process. I'll look into it some more. It's possible that presubmit_support itself is calling svn? I don't think CQ is closing stdin on presubmit. You're right, it should do that. On Tue, Jun 25, 2013 at 11:03 AM, Mark Seaborn <mseaborn@chromium.org>wrote: > On 25 June 2013 10:58, Robert Iannucci <iannucci@chromium.org> wrote: > >> As I commented on the other thread, the presubmit check is running a >> remote svn info command, which will not work on the CQ as it currently >> exists. The command hangs on a prompt for credentials. >> > > Which other thread was that? > > Would you happen to know what part of the code is calling "svn info", or > do you only know this based on the prompt you're seeing? I had a quick > look at NaCl's PRESUBMIT.py and tools/code_hygiene.py and didn't see any > obvious svn invocations. > > BTW, does the CQ redirect stdin to /dev/null? Would that help make this > fail fast rather than hang? > > Cheers, > Mark > > >> On Tue, Jun 25, 2013 at 10:51 AM, <commit-bot@chromium.org> wrote: >> >>> Presubmit check for 16880002-12001 failed and returned exit status -2001. >>> The presubmit check was hung. It took 360.0 seconds to execute and the >>> time >>> limit is 360.0 seconds. >>> >>> https://chromiumcodereview.appspot.com/16880002/ >>> >> >> > -- You received this message because you are subscribed to the Google Groups "Native-Client-Reviews" group. To unsubscribe from this group and stop receiving emails from it, send an email to native-client-reviews+unsubscribe@googlegroups.com. To post to this group, send email to native-client-reviews@googlegroups.com. Visit this group at http://groups.google.com/group/native-client-reviews. For more options, visit https://groups.google.com/groups/opt_out.
Hm... It looks like it may try to do a remote info call if it feels the need to generate a diff, which it'll do if anything in presubmit looks at ChangedContents or GenerateScmDiff. On Tue, Jun 25, 2013 at 11:10 AM, Robert Iannucci <iannucci@chromium.org>wrote: > I'm not sure, but I see presubmit_support.py with a svn info child > process. I'll look into it some more. It's possible that presubmit_support > itself is calling svn? > > I don't think CQ is closing stdin on presubmit. You're right, it should do > that. > > > On Tue, Jun 25, 2013 at 11:03 AM, Mark Seaborn <mseaborn@chromium.org>wrote: > >> On 25 June 2013 10:58, Robert Iannucci <iannucci@chromium.org> wrote: >> >>> As I commented on the other thread, the presubmit check is running a >>> remote svn info command, which will not work on the CQ as it currently >>> exists. The command hangs on a prompt for credentials. >>> >> >> Which other thread was that? >> >> Would you happen to know what part of the code is calling "svn info", or >> do you only know this based on the prompt you're seeing? I had a quick >> look at NaCl's PRESUBMIT.py and tools/code_hygiene.py and didn't see any >> obvious svn invocations. >> >> BTW, does the CQ redirect stdin to /dev/null? Would that help make this >> fail fast rather than hang? >> >> Cheers, >> Mark >> >> >>> On Tue, Jun 25, 2013 at 10:51 AM, <commit-bot@chromium.org> wrote: >>> >>>> Presubmit check for 16880002-12001 failed and returned exit status >>>> -2001. >>>> The presubmit check was hung. It took 360.0 seconds to execute and the >>>> time >>>> limit is 360.0 seconds. >>>> >>>> https://chromiumcodereview.appspot.com/16880002/ >>>> >>> >>> >> > -- You received this message because you are subscribed to the Google Groups "Native-Client-Reviews" group. To unsubscribe from this group and stop receiving emails from it, send an email to native-client-reviews+unsubscribe@googlegroups.com. To post to this group, send email to native-client-reviews@googlegroups.com. Visit this group at http://groups.google.com/group/native-client-reviews. For more options, visit https://groups.google.com/groups/opt_out.
**needs to generate a diff and there are renamed/moved files On Tue, Jun 25, 2013 at 11:20 AM, Robert Iannucci <iannucci@chromium.org>wrote: > Hm... It looks like it may try to do a remote info call if it feels the > need to generate a diff, which it'll do if anything in presubmit looks at > ChangedContents or GenerateScmDiff. > > > On Tue, Jun 25, 2013 at 11:10 AM, Robert Iannucci <iannucci@chromium.org>wrote: > >> I'm not sure, but I see presubmit_support.py with a svn info child >> process. I'll look into it some more. It's possible that presubmit_support >> itself is calling svn? >> >> I don't think CQ is closing stdin on presubmit. You're right, it should >> do that. >> >> >> On Tue, Jun 25, 2013 at 11:03 AM, Mark Seaborn <mseaborn@chromium.org>wrote: >> >>> On 25 June 2013 10:58, Robert Iannucci <iannucci@chromium.org> wrote: >>> >>>> As I commented on the other thread, the presubmit check is running a >>>> remote svn info command, which will not work on the CQ as it currently >>>> exists. The command hangs on a prompt for credentials. >>>> >>> >>> Which other thread was that? >>> >>> Would you happen to know what part of the code is calling "svn info", or >>> do you only know this based on the prompt you're seeing? I had a quick >>> look at NaCl's PRESUBMIT.py and tools/code_hygiene.py and didn't see any >>> obvious svn invocations. >>> >>> BTW, does the CQ redirect stdin to /dev/null? Would that help make this >>> fail fast rather than hang? >>> >>> Cheers, >>> Mark >>> >>> >>>> On Tue, Jun 25, 2013 at 10:51 AM, <commit-bot@chromium.org> wrote: >>>> >>>>> Presubmit check for 16880002-12001 failed and returned exit status >>>>> -2001. >>>>> The presubmit check was hung. It took 360.0 seconds to execute and the >>>>> time >>>>> limit is 360.0 seconds. >>>>> >>>>> https://chromiumcodereview.appspot.com/16880002/ >>>>> >>>> >>>> >>> >> > -- You received this message because you are subscribed to the Google Groups "Native-Client-Reviews" group. To unsubscribe from this group and stop receiving emails from it, send an email to native-client-reviews+unsubscribe@googlegroups.com. To post to this group, send email to native-client-reviews@googlegroups.com. Visit this group at http://groups.google.com/group/native-client-reviews. For more options, visit https://groups.google.com/groups/opt_out.
CQ is trying da patch. Follow status at https://nativeclient-status.appspot.com/cq/petarj@mips.com/16880002/12001
Presubmit check for 16880002-12001 failed and returned exit status -2001. The presubmit check was hung. It took 360.0 seconds to execute and the time limit is 360.0 seconds.
On 2013/06/27 00:21:16, I haz the power (commit-bot) wrote: > Presubmit check for 16880002-12001 failed and returned exit status -2001. > The presubmit check was hung. It took 360.0 seconds to execute and the time > limit is 360.0 seconds. Mark, how to submit this patch? Petar
CQ is trying da patch. Follow status at https://nativeclient-status.appspot.com/cq/petarj@mips.com/16880002/12001
Presubmit check for 16880002-12001 failed and returned exit status -2001. The presubmit check was hung. It took 360.0 seconds to execute and the time limit is 360.0 seconds.
CQ is trying da patch. Follow status at https://nativeclient-status.appspot.com/cq/petarj@mips.com/16880002/12001
Presubmit check for 16880002-12001 failed and returned exit status -2001. The presubmit check was hung. It took 360.0 seconds to execute and the time limit is 360.0 seconds.
On 2013/06/27 18:14:02, I haz the power (commit-bot) wrote: > Presubmit check for 16880002-12001 failed and returned exit status -2001. > The presubmit check was hung. It took 360.0 seconds to execute and the time > limit is 360.0 seconds. I'm pretty sure this will need to be committed manually. CQ can't run the presubmit for this because of the issues I mentioned previously.
On 2013/06/27 19:12:05, iannucci wrote: > On 2013/06/27 18:14:02, I haz the power (commit-bot) wrote: > > Presubmit check for 16880002-12001 failed and returned exit status -2001. > > The presubmit check was hung. It took 360.0 seconds to execute and the time > > limit is 360.0 seconds. > > I'm pretty sure this will need to be committed manually. CQ can't run the > presubmit for this because of the issues I mentioned previously. On 'git cl push' I get this error: error: Cannot access URL http://git.chromium.org/native_client/src/native_client.git/, return code 22 fatal: git-http-push failed Can you suggest how to proceed? Petar
On 27 June 2013 17:44, <petarj@mips.com> wrote: > On 2013/06/27 19:12:05, iannucci wrote: > >> On 2013/06/27 18:14:02, I haz the power (commit-bot) wrote: >> > Presubmit check for 16880002-12001 failed and returned exit status >> -2001. >> > The presubmit check was hung. It took 360.0 seconds to execute and the >> time >> > limit is 360.0 seconds. >> > > I'm pretty sure this will need to be committed manually. CQ can't run the >> presubmit for this because of the issues I mentioned previously. >> > > On 'git cl push' I get this error: > > error: Cannot access URL > http://git.chromium.org/**native_client/src/native_**client.git/<http://git.c..., > return code 22 > fatal: git-http-push failed > > Can you suggest how to proceed? > The native_client repo is in SVN, so you would have to use "git cl dcommit" to commit. "git cl push" is for Git repos only. (native_client.git above is just a mirror of SVN.) We'll have to give you SVN commit access for dcommit to work for you. I think we should do that, but I'll have to ask someone with admin rights about that tomorrow. We should also fix the CQ so that it works again. I haven't had time in the last couple of days to investigate why it has stopped working for file renames -- sorry. Mark -- You received this message because you are subscribed to the Google Groups "Native-Client-Reviews" group. To unsubscribe from this group and stop receiving emails from it, send an email to native-client-reviews+unsubscribe@googlegroups.com. To post to this group, send email to native-client-reviews@googlegroups.com. Visit this group at http://groups.google.com/group/native-client-reviews. For more options, visit https://groups.google.com/groups/opt_out.
On 2013/06/28 01:36:18, Mark Seaborn wrote: > On 27 June 2013 17:44, <mailto:petarj@mips.com> wrote: > > > On 2013/06/27 19:12:05, iannucci wrote: > > > >> On 2013/06/27 18:14:02, I haz the power (commit-bot) wrote: > >> > Presubmit check for 16880002-12001 failed and returned exit status > >> -2001. > >> > The presubmit check was hung. It took 360.0 seconds to execute and the > >> time > >> > limit is 360.0 seconds. > >> > > > > I'm pretty sure this will need to be committed manually. CQ can't run the > >> presubmit for this because of the issues I mentioned previously. > >> > > > > On 'git cl push' I get this error: > > > > error: Cannot access URL > > > http://git.chromium.org/**native_client/src/native_**client.git/%3Chttp://git...>, > > return code 22 > > fatal: git-http-push failed > > > > Can you suggest how to proceed? > > > > The native_client repo is in SVN, so you would have to use "git cl dcommit" > to commit. "git cl push" is for Git repos only. (native_client.git above > is just a mirror of SVN.) We'll have to give you SVN commit access for > dcommit to work for you. I think we should do that, but I'll have to ask > someone with admin rights about that tomorrow. > > We should also fix the CQ so that it works again. I haven't had time in > the last couple of days to investigate why it has stopped working for file > renames -- sorry. > Mark, any news here? Petar
Message was sent while issue was closed.
Committed patchset #3 manually as r11674 (presubmit successful).
On 3 July 2013 12:18, <mseaborn@chromium.org> wrote: > Committed patchset #3 manually as r11674 (presubmit successful). > I've committed the change on your behalf. Sorry that committing this change has been so painfully slow. I'm not sure if the commit queue is working for NaCl at all at the moment, or if the problem only occurred because the change contained file renames. If you have more problems with the commit queue, can you file an issue via https://code.google.com/p/chromium/issues/entry?template=Build%20Infrastructu... CC me, please? Cheers, Mark -- You received this message because you are subscribed to the Google Groups "Native-Client-Reviews" group. To unsubscribe from this group and stop receiving emails from it, send an email to native-client-reviews+unsubscribe@googlegroups.com. To post to this group, send email to native-client-reviews@googlegroups.com. Visit this group at http://groups.google.com/group/native-client-reviews. For more options, visit https://groups.google.com/groups/opt_out. |